Aerobiological Stabilities of Different Species of Gram-Negative Bacteria, Including Well-Known Biothreat Simulants, in Single-Cell Particles and Cell Clusters of Different Compositions

ABSTRACT

The ability to perform controlled experiments with bioaerosols is a fundamental enabler of many bioaerosol research disciplines. A practical alternative to using hazardous biothreat agents, e.g., for detection equipment development and testing, involves using appropriate model organisms (simulants). Several species of Gram-negative bacteria have been used or proposed as biothreat simulants. However, the appropriateness of different bacterial genera, species, and strains as simulants is still debated. Here, we report aerobiological stability characteristics of four species of Gram-negative bacteria (Pantoea agglomerans, Serratia marcescens, Escherichia coli, and Xanthomonas arboricola) in single-cell particles and cell clusters produced using four spray liquids (H₂O, phosphate-buffered saline[PBS], spent culture medium[SCM], and a SCM-PBS mixture). E. coli showed higher stability in cell clusters from all spray liquids than the other species, but it showed similar or lower stability in single-cell particles. The overall stability was higher in cell clusters than in single-cell particles. The highest overall stability was observed for bioaerosols produced using SCM-containing spray liquids. A key finding was the observation that stability differences caused by particle size or compositional changes frequently followed species-specific patterns. The results highlight how even moderate changes to one experimental parameter, e.g., bacterial species, spray liquid, or particle size, can strongly affect the aerobiological stability of Gram-negative bacteria. Taken together, the results highlight the importance of careful and informed selection of Gram-negative bacterial biothreat simulants and also the accompanying particle size and composition. The outcome of this work contributes to improved selection of simulants, spray liquids, and particle size for use in bioaerosol research.

IMPORTANCE The outcome of this work contributes to improved selection of simulants, spray liquids, and particle size for use in bioaerosol research. Taken together, the results highlight the importance of careful and informed selection of Gram-negative bacterial biothreat simulants and also the accompanying particle size and composition. The results highlight how even moderate changes to one experimental parameter, e.g., bacterial species, spray liquid, or particle size, can strongly affect the aerobiological stability of Gram-negative bacteria. A key finding was the observation that stability differences caused by particle size or compositional changes frequently followed species-specific patterns.

INTRODUCTION

Biological aerosols (bioaerosols) have been studied in indoor and outdoor environments for more than a century and for several different reasons, including biodefense, atmospheric science, indoor air quality control, and microbial ecology-related questions (1). Bioaerosol research is inherently multidisciplinary and demands scientists with an integrated skill set, including microbiology, molecular biology, occupational and public health, and aerosol physics (2). Exposure to bioaerosols may result in a range of adverse health effects, including infectious disease, acute toxic effects, and allergy (3). The importance of bioaerosol research has in recent years been highlighted by several infectious disease outbreaks as well as the emerging bioterrorism threat capitalized by the 2001 Amerithrax incidents in the United States (2).

The ability to perform controlled experimental studies with bioaerosols is a fundamental enabler in many bioaerosol research disciplines, including efforts to address their properties, sources, and fate. Development and testing of detection systems for biothreat agents is a branch of aerosol science that has received increased attention within the defense and security domain in recent years. Due to their hazardous nature, most or all biothreat agents demand strict containment measures even for basic handling, and bioaerosol research activities involving biothreat agents are therefore by default complex and expensive endeavors restricted to specialized containment facilities. In addition to the technical, practical, and even legal aspects of working with biothreat agents in controlled laboratory environments, bioaerosol research often involves scientific questions which can be properly addressed only under realistic atmospheric conditions.

A practical alternative to using hazardous biothreat agents, e.g., for development and testing of biothreat detection systems, involves the use of appropriate model organisms. Model organisms used for development and testing purposes within the defense and security domain are generally referred to as biothreat simulants. Biothreat simulants are microorganisms or biological materials that are intended to mimic certain characteristic properties of biothreat agents but without, or at least with significantly reduced, associated health hazard potential. Biothreat simulants typically are environmental microorganisms that have some physical, chemical, or biological properties in common with at least one biothreat agent. The use of simulants instead of biothreat agents has obvious benefits in terms of reducing the health, safety, security, and environmental concerns associated with bioaerosol research as well as the added benefit of allowing bioaerosol research, including development and testing of detection systems, to be performed outside specialized containment facilities.

Simulants have been used for almost as long as the biothreat agents themselves, and one of the best known biothreat simulants is Bacillus atrophaeus, formerly Bacillus globigii, which has been used as a simulant for Bacillus anthracis spores for more than half a century (4). Four main classes of simulants are used in bioaerosol research: (i) spores of Gram-positive bacteria (e.g., B. atrophaeus spores) as a surrogate for Bacillus anthracis, (ii) vegetative cells of Gram-negative bacteria (e.g., Pantoea agglomerans, formerly Erwinia herbicola) as a surrogate for, e.g., Yersinia pestis, (iii) viruses (e.g., bacteriophage MS2) as a surrogate for, e.g., variola virus, and (iv) proteins (e.g., ovalbumin) as a surrogate for, e.g., botulinum toxin (5).

Several species of Gram-negative bacteria have been used or proposed as biothreat simulants or general model organisms for Gram-negative bacteria, including P. agglomerans (6), Serratia marcescens (7, 8), Escherichia coli (9), and Xanthomonas arboricola (10). However, the appropriateness of different Gram-negative bacterial genera, species, and strains as biothreat simulants is still an ongoing debate (5). Another complicating factor is the fact that in addition to how well simulants mimic relevant properties of biothreat agents, which often depend on the specific context of use, several other technical, practical, economic, and even legal aspects have to be taken into account as part of simulant selection processes.

In many situations, the aerobiological stability of biothreat agents, i.e., the ability to survive and remain viable/infectious in an airborne state, will determine the associated health hazard potential, since as long as the biomaterial has been rendered nonviable/noninfectious, inhalation of even large amounts of a highly pathogenic microorganism represents no infectious disease risk (11). Knowledge about the aerobiological stability of biothreat agents therefore is fundamental to correctly assess the potential health implications following an aerosol attack involving inhalation exposure to biothreat agents (11). In response to a range of technical and practical challenges associated with suspending microorganisms as an aerosol for extended periods of time under controlled experimental conditions, several practical approaches have been developed to study aerobiological stability of microorganisms, including stirred-settling aerosol chambers (12), rotating aerosol drums (13), and microthread-based suspension techniques (14).

The aerobiological stability of several species of Gram-negative bacteria, including biothreat agents such as Y. pestis and Francisella tularensis and simulants such as P. agglomerans and S. marcescens, have previously been studied by several investigators (11, 15, 16). The atmosphere is generally considered a hostile environment for microorganisms, and a range of environmental conditions play a role in governing the aerobiological stability of Gram-negative bacteria, including temperature, relative humidity, electromagnetic radiation, the open-air factor, and atmospheric oxygen content (11, 15, 16). In addition to environmental conditions, several other factors have also been shown to influence the aerobiological stability of Gram-negative bacteria, including type of aerosol generation and sampling methods, formulation of spray and collection liquids, growth phase, medium composition, culture conditions, and physiological age of the microbial culture (11, 15, 16).

The aim of this study was to investigate and directly compare the aerobiological stability characteristics of four different species of Gram-negative bacteria that previously have been proposed or used as biothreat simulants or general model organisms in bioaerosol research. Vegetative cells of P. agglomerans, S. marcescens, E. coli, and X. arboricola were each aerosolized from four different spray liquid formulations, namely, ultrapure water (H₂O), phosphate-buffered saline (PBS), spent culture medium (SCM), and an SCM-PBS mixture, as single-cell (∼1 μm) and cell cluster (∼3.5 μm) droplet nucleus particles. The comprehensive experimental data set was used to highlight differences and similarities in the aerobiological stability of different Gram-negative bacterial species as well as the relative influence of other experimental parameters, including spray liquid composition and particle size. Taken together, this work has relevance for bioaerosol research, including public health- and biodefense-related applications. The outcome of this study may contribute to improved selection of biothreat simulants as well as better-informed selection of accompanying spray liquid formulations and particle sizes for use in different bioaerosol research applications.

RESULTS

Size distribution of single-cell particles and cell clusters of different composition.

The mass median aerodynamic diameters and geometric standard deviations of the single-cell (∼1 μm) and cell cluster (∼3.5 μm) droplet nuclei were stable and reproducible throughout the study (Table 1). Since the concentration of bacterial cells were kept at a constant level in the spray liquids, i.e., 4 × 10⁷ ± 1 × 10⁷ CFU ml⁻¹ for single-cell particles and 4 × 10⁸ ± 1 × 10⁸ CFU ml⁻¹ for cell clusters, the mass median aerodynamic diameter and geometric standard deviation was mainly determined by the droplet size distribution of the aerosol generator and the spray liquid composition (Table 1). The droplet nuclei produced from spray liquids containing PBS, SCM, and SCM-PBS were larger than those produced from H₂O (Table 1). While this effect was observed for both aerosol generators, it was more pronounced for the ultrasonic atomizer than the nebulizer (Table 1). Since the liquid droplets produced by the ultrasonic atomizer (∼38 μm) were larger than those produced by the nebulizer (<10 μm), this observation was not completely unexpected; however, it is still noteworthy, since diluted spray liquids, i.e., PBSd (0.025× PBS diluted in H₂O), SCMd (0.2× SCM diluted in H₂O), and SCMd-PBSd, were used together with the ultrasonic atomizer in an attempt to compensate for the known droplet size difference. At the fixed-concentration level of bacterial cells used in this work, the droplet nuclei produced with the nebulizer were shown to consist mainly of individual cells or spores, while those produced with the ultrasonic atomizer were shown to mainly consist of cell or spore clusters (Fig. 1).

TABLE 1.

Particle size distributions for aerosolized droplet nuclei containing bacterial cells or spores produced using different aerosol generators and spray liquid formulations

Organism	Particle size distributiona
	Mean (±SD) mass median aerodynamic diameter (μm)								Mean (±SD) geometric standard deviation
	48kHz Sono-Tek ultrasonic atomizer				Hudson Micro Mist nebulizer				48kHz Sono-Tek ultrasonic atomizer				Hudson Micro Mist nebulizer
	H2O	PBSd	SCMd	SCMd-PBSd	H2O	PBS	SCM	SCM-PBS	H2O	PBSd	SCMd	SCMd-PBSd	H2O	PBS	SCM	SCM-PBS
Pantoea agglomerans	2.61 ± 0.06	3.06 ± 0.06	3.95 ± 0.12	4.29 ± 0.22	0.79 ± 0.02	1.09 ± 0.06	1.03 ± 0.01	1.08 ± 0.05	1.19 ± 0.02	1.17 ± 0.00	1.17 ± 0.00	1.17 ± 0.00	1.08 ± 0.01	1.31 ± 0.06	1.26 ± 0.01	1.32 ± 0.01
Serratia marcescens	2.97 ± 0.16	3.28 ± 0.14	3.84 ± 0.05	4.05 ± 0.01	0.68 ± 0.00	1.33 ± 0.04	1.05 ± 0.00	1.25 ± 0.02	1.18 ± 0.02	1.18 ± 0.01	1.17 ± 0.00	1.17 ± 0.01	1.07 ± 0.00	1.34 ± 0.01	1.24 ± 0.00	1.34 ± 0.02
Escherichia coli	2.78 ± 0.03	3.06 ± 0.07	3.86 ± 0.03	4.06 ± 0.01	0.71 ± 0.01	1.13 ± 0.01	1.02 ± 0.07	1.06 ± 0.02	1.19 ± 0.02	1.18 ± 0.01	1.19 ± 0.01	1.18 ± 0.00	1.09 ± 0.01	1.34 ± 0.00	1.23 ± 0.01	1.33 ± 0.00
Xanthomonas arboricola	2.88 ± 0.19	3.22 ± 0.17	3.95 ± 0.11	4.13 ± 0.11	0.71 ± 0.01	1.22 ± 0.02	1.02 ± 0.07	1.20 ± 0.02	1.18 ± 0.02	1.17 ± 0.01	1.17 ± 0.01	1.18 ± 0.01	1.07 ± 0.00	1.37 ± 0.02	1.28 ± 0.01	1.35 ± 0.01
Bacillus atrophaeus spores	2.90 ± 0.04	3.31 ± 0.06	3.88 ± 0.06	4.24 ± 0.06	0.92 ± 0.03	1.19 ± 0.04	1.16 ± 0.03	1.21 ± 0.02	1.19 ± 0.00	1.21 ± 0.00	1.19 ± 0.01	1.20 ± 0.01	1.18 ± 0.04	1.39 ± 0.01	1.36 ± 0.01	1.38 ± 0.01

^aParticle size distribution was calculated based on APS 3321 aerodynamic particle sizer data from five or more experiments for each organism.

FIG 1.

Scanning electron microscopy (SEM) images showing droplet nuclei containing Bacillus atrophaeus spores in spore clusters (bottom and top left) and single-spore particles (bottom and top right). Aerosolized droplet nuclei were sampled onto polycarbonate membrane filters (0.4-μm pore size), sputter coated (2 nm Pt/Pd; Cressington 208HR), and imaged (high-vacuum mode, 1 kV; Hitachi SU6600 FE-SEM).

Normalization of aerobiological stability data.

Physical particle loss occurring in the aerosol chamber during the aerobiological stability experiments, e.g., due to gravitational settling, inertial impaction onto fan blades, and electrostatic interactions, had to be accounted for to specifically quantify the aerobiological decay. Therefore, APS 3321-recorded particle counting and sizing data were used to establish physical decay rates for single-cell particles and cell clusters containing P. agglomerans, S. marcescens, E. coli, or X. arboricola cells, or B. atrophaeus spores, in combination with each spray liquid. B. atrophaeus spores were used as a reference because no measurable aerobiological decay occurred during the investigated aerosol residence time (data not shown). Aerosol experiments with B. atrophaeus spores showed that the physical decay rates based on APS 3321 data corresponded well with the aerobiological decay rates based on STA-203 data for droplet nuclei generated with both aerosol generators and from all spray liquids (data not shown). The APS 3321-derived physical decay rates were similar for same-sized droplet nuclei irrespective of whether the payload was P. agglomerans, S. marcescens, E. coli, or X. arboricola cells or B. atrophaeus spores (data not shown). In addition to facilitating data normalization, parallel aerosol experiments with B. atrophaeus spores also enabled expression of the aerobiological decay rates for Gram-negative bacteria relative to B. atrophaeus spores under similar experimental conditions. In turn, this allowed for direct comparison of the aerobiological stability characteristics for different Gram-negative bacteria aerosolized using different spray liquids and aerosol generators, which otherwise would not have been possible due to inherent differences in the resulting aerosol size, composition, generation mechanism, and flux rate.

Aerobiological stability of Gram-negative bacteria in single-cell particles and cell clusters of different composition.

The aerobiological stability of four different species of Gram-negative bacteria (P. agglomerans, S. marcescens, E. coli, and X. arboricola) in single-cell (∼1 μm) and cell cluster (∼3.5 μm) droplet nuclei produced from four different spray liquids (H₂O, PBS, SCM, and SCM-PBS) were established based on five or more aerosol experiments for each organism (Fig. 2; also see Table S1 in the supplemental material). The results were expressed as percent residual survival at different time points (0 min, 5 min, 10 min, and 20 min) relative to that observed for a stable reference (B. atrophaeus spores) under similar experimental conditions (Fig. 2; Table S1). The aerosol residence time needed for inactivation of 90% of the particle population (d90) was calculated using standard exponential, logarithmic, or linear best-fit regression analysis (Table 2).

FIG 2.

Aerobiological stabilities of different species of Gram-negative bacteria in single-cell particles produced with a nebulizer (∼1 μm; right) and cell clusters produced with an ultrasonic atomizer (∼3.5 μm; left) from different spray liquids. Diluted solutions (diluted with H₂O) of PBS, SCM-PBS, and SCM were used together with the ultrasonic atomizer, while undiluted solutions were used together with the nebulizer. Aerobiological decay rates were calculated based on five or more aerosol experiments for each organism and are reported as percent residual survival (means ± standard deviations) at different time points relative to that observed for a stable reference (B. atrophaeus spores) under similar experimental conditions.

TABLE 2.

Aerobiological stabilities of different species of Gram-negative bacteria in single-cell particles and cell clusters produced from different spray liquids

Organism	Aerobiological decay rate (d90, min) ina:
	H2O	PBS (PBSd)	SCM-PBS (SCMd-PBSd)	SCM (SCMd)
Single-cell droplet nuclei (∼1 μm)
Pantoea agglomerans	<1	<1	4	9
Serratia marcescens	<1	3	16	<1
Escherichia coli	<1	<1	14	<1
Xanthomonas arboricola	<1	<1	2	6
Cell cluster droplet nuclei (∼3.5 μm)
Pantoea agglomerans	9	<1	5	2
Serratia marcescens	<1	<1	11	4
Escherichia coli	108	48	287	143
Xanthomonas arboricola	4	2	2	3

^aDiluted solutions (diluted with H₂O) of PBS, SCM-PBS, and SCM were used together with the ultrasonic atomizer, while undiluted solutions were used together with the nebulizer. Aerobiological decay rates were calculated based on five or more aerosol experiments for each organism and are reported as the aerosol residence time (in minutes) needed to inactivate 90% of the particle population (d90). Standard exponential, logarithmic, or linear best-fit regression was used to calculate d90 values. If d90 occurred before 1 min, the d90 value is reported as <1 min. Boldface indicates a rate of ≥10 min, italics indicates a rate of ≥1 min, and normal font indicates a rate of <1 min.

The results revealed significant differences between the studied Gram-negative bacterial species regarding their aerobiological stability (Fig. 2 and Table 2; Table S1). E. coli was shown to have significantly higher stability (d90, ≥48 min) in cell clusters from all spray liquids than P. agglomerans, S. marcescens, and X. arboricola, which all had a d90 of ≤1 min (Table 2). However, similar results were not observed for single-cell particles where E. coli had similar or significantly lower stability than P. agglomerans, S. marcescens, and X. arboricola from all spray liquids (Table 2).

Significant differences were also observed between the aerobiological stability characteristics of Gram-negative bacteria in single-cell particles and those in cell clusters (Fig. 2 and Table 2; Table S1). The stability was found to be similar or significantly lower in single-cell particles than in cell clusters for most combinations of bacterial species and spray liquid formulations (Table 2). The only exceptions were S. marcescens sprayed from PBS, which had a d90 of ∼3 min in single-cell particles compared to <1 min in cell clusters, S. marcescens from SCM-PBS, with a d90 of ∼16 min in single-cell particles compared to ∼11 min in cell clusters, P. agglomerans from SCM, with a d90 of ∼9 min in single-cell particles compared to ∼2 min in cell clusters, and X. arboricola from SCM, with a d90 of ∼6 min in single-cell particles compared to ∼3 min in cell clusters (Table 2).

The aerobiological stability of Gram-negative bacteria was in many cases significantly different in similar-sized droplet nuclei of different compositions (Fig. 2 and Table 2; Table S1). All of the Gram-negative bacteria tested (P. agglomerans, S. marcescens, E. coli, and X. arboricola) had very low stability (d90, <1 min) in single-cell particles when H₂O was used as a spray liquid (Table 2). Very low stability (d90, <1 min) was also observed for P. agglomerans, E. coli, and X. arboricola from PBS and S. marcescens and E. coli from SCM in single-cell particles and for S. marcescens from H₂O and P. agglomerans and S. marcescens from PBSd in cell clusters (Table 2). The highest stability encountered in this work was observed for E. coli in cell clusters produced from SCMd-PBSd with a d90 of ∼287 min, while the second highest stability was observed for E. coli in cell clusters produced from SCMd with a d90 of ∼143 min (Table 2). However, the highest stability in single-cell particles was observed for S. marcescens disseminated from SCM-PBS with a d90 of ∼16 min, followed by E. coli disseminated from SCM-PBS with a d90 of ∼14 min (Table 2).

DISCUSSION

Here, we report aerobiological stability characteristics of four different species of Gram-negative bacteria (P. agglomerans, S. marcescens, E. coli, and X. arboricola) which previously have been proposed or used as biothreat simulants or general model organisms in bioaerosol research. A comprehensive experimental data set was generated based on more than 200 aerosol experiments to allow for direct comparison of the aerobiological stability characteristics of different species of Gram-negative bacteria in single-cell particles and cell clusters produced from different spray liquids (H₂O, PBS, SCM, and SCM-PBS).

Although the aerobiological stability characteristics of several species of Gram-negative bacteria already have been investigated, including S. marcescens (7, 8, 17, 18), E. coli (19–25), and P. agglomerans (6, 12, 26–30), most of these studies were performed more than 3 decades ago; thus, several knowledge gaps still remain. An inherent challenge regarding the aerobiological stability of Gram-negative bacteria is the observation that few generalizations seem to be applicable, suggesting that different genera, species, and even strains of Gram-negative bacteria have to be considered separate entities regarding their stability characteristics (11, 15, 16). Adding to the challenge is the observation that previous studies have often reported variable and sometimes even conflicting results, which may be attributed to large heterogeneity regarding the scope, experimental design, and analytical methods involved in aerobiology research (11, 15, 16). Therefore, a substantial portion of the existing stability data for Gram-negative bacteria may be of limited value, especially for interstudy comparison and generalization purposes, since a range of experimental parameters which often have been insufficiently controlled or monitored may have acted in concert to alter the stability characteristics in a complex and largely unknown way (11, 15, 16). In a situation where it is difficult to control or monitor the full range of potentially confounding variables, a fruitful experimental approach may be to investigate the stability characteristics of different Gram-negative bacteria using the same experimental approach and conditions and thereby focus on intrastudy rather than interstudy result comparisons. Unfortunately, previous studies addressing the stability characteristics of Gram-negative bacteria most often have been restricted to the study of a single bacterial species. Therefore, limited information is available on relative differences in the stability of different Gram-negative bacteria, and the representativeness and, to some extent, the relevance of previous findings that have been recorded for only a single species is, in many cases, still largely unknown (6–8, 12, 17–30). This study allows for direct intrastudy comparison of the aerobiological stability characteristics of four different species of Gram-negative bacteria based on results obtained using the same experimental approach, setup, and conditions. To our knowledge, the stability of more than one of the four bacterial species studied in this work had not previously been investigated in the same study except for a few investigations involving both E. coli and S. marcescens (31–36); thus, the stabilities of P. agglomerans and X. arboricola have not been directly compared with that of E. coli or S. marcescens. Most investigations of Gram-negative bacteria to date have also involved only a single spray liquid formulation and particle size, which adds to the challenge of comparing and extrapolating results between studies, since differences in at least one experimental parameter that may play an important role in governing the aerobiological stability of Gram-negative bacteria, e.g., bacterial species, spray liquid formulation, particle size, or environmental conditions, usually exist (11, 15, 16).

E. coli was shown to have significantly higher aerobiological stability, i.e., a slower aerobiological decay rate, than P. agglomerans, S. marcescens, and X. arboricola in cell clusters produced from all spray liquids tested (Fig. 2 and Table 2; also see Table S1 in the supplemental material). This finding is in general agreement with previous investigations, which have often reported relatively high stability for E. coli (25, 34, 37). In a previous study of the same strain of E. coli as that involved in this work (MRE162), a d90 value of ∼264 min was observed when E. coli was disseminated from SCM (25), which corresponds well with the d90 values of ∼143 min and ∼287 min observed in this work when E. coli was disseminated from SCM and SCM-PBS, respectively (Table 2).

Several of the most important and informative findings from this study would not have been seen if only a single species, spray liquid, or particle size had been investigated. As already mentioned, the stability of Gram-negative bacteria has been shown to be genus, species, or even strain specific (11, 15, 16, 25, 34, 37). A species-specific finding from the current work in support of this was the observation that while E. coli showed higher stability in cell clusters than P. agglomerans, S. marcescens, and X. arboricola, in single-cell particles E. coli showed similar and, in many cases, even lower stability (Table 2). When the collective results for both single-cell particles and cell clusters and all species of bacteria tested are considered, the highest overall stability was observed for droplet nuclei generated from spray liquids containing SCM (Fig. 2 and Table 2; Table S1). This finding is in agreement with previous investigations that have reported higher stability of Gram-negative bacteria aerosolized from SCM-containing spray liquids compared to those containing H₂O and PBS (30). In all but a few cases the stability of Gram-negative bacteria was higher in cell clusters than in single-cell particles in this study (Fig. 2 and Table 2; Table S1), which is in general agreement with the results from previous investigations (14, 25, 28, 30, 38).

Taken together, the reported results have clearly shown how even modest changes to one experimental parameter, e.g., choice of bacteria, spray liquids, or particle sizes, can have a major impact on the aerobiological stability of the resulting bioaerosol. Although some general trends regarding the stability of Gram-negative bacteria may be extrapolated and generalized from the results, a key finding was the observation that stability differences which could be attributed to changes in particle size (e.g., due to different droplet sizes) or composition (e.g., due to different spray liquid formulations) in most cases followed a species-specific pattern (Fig. 2 and Table 2; Table S1). This finding lends further support to the concept that generalization and extrapolation of the aerobiological stability characteristics for Gram-negative bacteria are complex tasks which may also be of limited relevance and value because a substantial portion of the characteristics may be genus, species, or even strain specific (11, 15, 16, 25, 34, 37).

Our findings highlight the importance of careful and informed selection of Gram-negative bacteria for use as biothreat simulants as well as the potential need for considering application-specific biothreat simulant selection processes, since a single Gram-negative bacterial simulant may not necessarily be optimal or even suitable for all applications. In combination with similar aerobiological stability data for Gram-negative bacterial biothreat agents, the results from this study could be used to guide and improve the selection of biothreat simulants for use in different bioaerosol research applications, including development and testing of biothreat detection systems. However, limited information is available in the literature on the aerobiological stability of Gram-negative bacterial biothreat agents (39–44). Our findings also highlight the importance of careful and informed selection of spray liquids and particle size, since the results showed that these factors may have a similar and even higher impact on the resulting aerobiological stability than the choice of Gram-negative bacteria.

Although this work involved several species of Gram-negative bacteria, spray liquids, and particle sizes, many environmental conditions and experimental variables which also warrant study were not investigated. Therefore, it may be interesting to launch follow-up studies that use the same test environment and experimental setup to investigate the influence of additional environmental conditions and experimental variables, including temperature, relative humidity, the open-air factor, solar radiation, culture conditions, culture age, and spray liquid additives (e.g., osmoprotectants). It also could be interesting to study several strains of the same bacterial species to specifically address strain-specific aerobiological stability characteristics. The ambition for the future also involves adopting additional characterization methods, including flow cytometry in combination with membrane integrity and metabolic cell staining (20, 45, 46) and quantitative PCR in combination with ethidium or propidium monoazide live-dead discrimination techniques (47–50), as well as the study of multiple strains of the same bacterial species to investigate strain-specific aerobiological stability characteristics.

MATERIALS AND METHODS

Test material.

Four different species of Gram-negative bacteria were used: P. agglomerans (ATCC 33243), S. marcescens (ATCC 274), E. coli (MRE162), and X. arboricola (P8-10). P. agglomerans and S. marcescens were purchased from the American Type Culture Collection (Manassas, VA), E. coli was kindly provided by the Defense Science and Technology Laboratory (Porton Down, United Kingdom), and X. arboricola was kindly provided by the Direction Générale de L'Armement (Vert-Le-Petit, France). Seed stocks were recovered from storage (−80°C) and inoculated into 100 ml nutrient broth (Merck, Darmstadt, Germany). The bacterial cultures were grown in a shaking incubator (200 rpm) at 30°C for 18 h (P. agglomerans and S. marcescens), 37°C for 18 h (E. coli), or 26°C for 42 h (X. arboricola).

Bioaerosol test facility.

Bioaerosol experiments were performed in a 12-m³ (3 by 2 by 2 m) aerosol chamber (Dycor Technologies, Edmonton, Canada) fitted with external heating, ventilation, and air conditioning and high-efficiency particulate air filtration of inlet and exhaust air. The chamber was equipped with internal mixing fans, temperature, humidity, and barometric pressure sensors, a Grimm 1.108 optical particle counter (Grimm Technologies, Douglasville, GA), an APS 3321 aerodynamic particle sizer (TSI, Shoreview, MN), and STA-203 slit-to-agar (STA) samplers (New Brunswick, Edison, NJ). Real-time monitoring of aerosol concentration and size distribution was done with Grimm 1.108 and APS 3321. Aerosol concentration levels of viable (cultivable) bacterial cells were determined using STA-203 samplers and expressed agent-containing particles per liter of air.

Spray liquid formulation.

Four different spray liquids were used together with each bacterial species to generate aerosolized droplet nuclei (∼1 μm) containing single cells. Briefly, aliquots (1 ml) of freshly prepared bacterial culture were harvested by centrifugation (4,000 × g, 10 min) and resuspended in 25 ml of H₂O (18.2 Ω cm⁻¹), PBS, sterile filtered SCM (0.2-μm pore size), or 1:4 SCM-PBS mixture. Four different spray liquids were used to generate aerosolized droplet nuclei (∼3.5 μm) containing cell clusters. Briefly, aliquots (10 ml) of freshly prepared bacterial culture were harvested by centrifugation (4,000 × g, 10 min) and resuspended in 25 ml of H₂O, 0.025× PBSd (PBS diluted in H₂O), 0.2× SCMd (SCM diluted in H₂O), or 1:4 SCMd-PBSd mixture. Diluted spray liquid formulations were used because the salt and organic material content (e.g., carbohydrates and cellular debris) was too high in undiluted PBS and SCM to be compatible with the droplet size of the aerosol generator that was used to produce cell clusters. The concentration level of bacterial cells in spray liquids was determined by plate count analysis and expressed as CFU per milliliter. Briefly, serial dilutions of spray liquids were plated in triplicate on nutrient agar (Merck) and incubated as previously described. The volume of bacterial culture harvested for each organism was adjusted to produce spray liquids containing 4 × 10⁷ ± 1 × 10⁷ CFU ml⁻¹ for single-cell particle generation and 4 × 10⁸ ± 1 × 10⁸ CFU ml⁻¹ for cell cluster generation. Fresh spray liquids were prepared for each experiment and used within 3 h.

Bioaerosol generation.

Droplet nuclei (∼1 μm) containing single bacterial cells were generated using a Hudson RCI Micro Mist nebulizer (Teleflex, Gurnee, IL) operated with ultrapure 5.0 N₂ gas (2.4 × 10⁴ pascals) producing droplets with a median diameter of <10 μm inside the aerosol chamber. Droplet nuclei (∼3.5 μm) containing cell clusters were generated using a 48-kHz Sono-Tek ultrasonic atomizer nozzle (Sono-Tek, Milton, NY) producing droplets with a median diameter of ∼38 μm. The Sono-Tek nozzle was powered by an ultrasonic generator (model 06-5108; Sono-Tek) operated at 6 W. Spray liquid was fed (800 μl min⁻¹) into the Sono-Tek using a syringe feeder (model 997E; Sono-Tek). The Sono-Tek was enclosed in an ADS-A20 aerosol dilution system (Dycor Technologies), providing a user-adjustable dilution of the aerosol with HEPA-filtered air before chamber injection. In order to maintain a static stirred-settling aerosol, the chamber was sealed by closing all ventilation dampers upon completion of aerosol generation and mixing (at 0 min). The aerosol was mixed by recirculation fans for the duration of the experiments. The chamber was operated at 20°C and 50% relative humidity, and these conditions remained stable for the duration of the experiments. Air sampling with STA-203 samplers for 2 min at 30 liters of air per min onto nutrient agar plates was initiated at 0, 5, 10, and 20 min. The STA-203 plates were incubated as described above. Colony counting was performed using a ProtoCOL HR automated colony counter (Synbiosis, Cambridge, United Kingdom), and the results are expressed as agent-containing particles per liter of air.

Normalization of aerobiological stability data using B. atrophaeus spores as a stable reference.

Parallel aerosol experiments were performed with B. atrophaeus spores as a stable reference to normalize the aerobiological stability data for P. agglomerans, S. marcescens, E. coli, and X. arboricola by expressing the results relative to those obtained for B. atrophaeus spores under similar experimental conditions. B. atrophaeus spores were obtained as a dry preparation (lot 19076-03268) from the U.S. Army Dugway Proving Ground (Dugway, UT). B. atrophaeus spores (100 mg) were suspended in H₂O (20 ml) and pelleted by centrifugation (10 min, 4,000 × g) after overnight storage in solution at 4°C. The supernatant was removed and the pellet resuspended in H₂O (20 ml). Spray liquids containing B. atrophaeus spores were prepared in accordance with the procedure for P. agglomerans. The plate count analysis and incubation procedure for B. atrophaeus spores was the same as that for P. agglomerans. The SCM used to prepare SCM-containing spray liquids with B. atrophaeus spores was derived from P. agglomerans cultures.

Statistical analysis.

The results were subjected to statistical analyses using SigmaPlot v12.3 (Systat Software, Inc., San Jose, CA). Normality testing was done with the Shapiro-Wilk test and, depending on whether the normality and equal variance criteria were fulfilled or not, significance testing was performed with Student's t test or Mann-Whitney rank-sum test. The significance level was set at a P value of <0.05 for all statistical tests.